每周论坛之一（总第258期）：柯紫筠（中山大学），2021年3月1日下午

每周论坛之一（总第258期）

报告人：柯紫筠 副教授

报告人单位：中山大学心理学系

报告主题：计量心理学

题目：Estimating Causal Effects in Meta-Analysis: The MASEM Approach Combined with the Instrumental Variables Regression Methods

时间：2021年3月1 日 （周一，14:20-16:00）

地点：中山大学东校园心理学系（南学院楼C座）305

报告简介：

When randomized experiments are impractical or immoral, the instrumental variables regression (IVR) methods are one of the viable methods for making causal inference. Previous studies on the IVR methods are largely limited to single empirical studies. We show that by implementing the IVR methods within the structural equation modeling (SEM) framework and by taking the advantage of the connection between SEM and meta-analytic SEM techniques (MASEM), we can extend the IVR methods to the context of meta-analysis. The extended IVR method does not require that each primary study investigates all relevant variables. We provide a tutorial on how to apply the extended IVR method using one of the MASEM technique, i.e., one-stage meta-analytic SEM (OSMASEM), on the aggregated level or equivalently with respect to the mean correlation matrix. We also provide an R function to examine the qualifying conditions for instrumental variables. We demonstrate the extended IVR method using an example on the relationship between innovation performance and business performance. Finally, we end the article with a discussion on the limitations and future directions of the extended method.

报告人简介：

柯紫筠，博士，中山大学心理学系副教授。研究兴趣为元分析技术、多层建模技术、稳健统计分析技术等。相关研究发表在Psychological Methods，Structural Equation Modeling，British Journal of Mathematical and Statistical Psychology等计量心理学领域知名国际期刊。